Latest page update: January 2022 (updated the CW-transmitter keying section)

Previous page updates: 30 November 2021 (added ref. 42); 22 October 2021 (expanded Fig 26 and dongle useability with a smartphone); 30 March 2021 (added Fig. 17, Fig. 18, Fig. 23 + text), November-December 2019 (added Fig. 9 & 10, added ref. 22D and associated Fig. 34)

©2004-2022 F. Dörenberg, unless stated otherwise. All rights reserved worldwide. No part of this publication may be used without permission from the author.

red-blue line


Rudolf Hell invented both the Hellschreiber and the digital "bitmap" font in 1929. It took another 50 years for Hellschreiber software to be developed. During the 1980s-1990s, a number of software emulators for Hellschreibers were developed for home computers such as PC (DOS, Windows®, LINUX), Apple/Macintosh, Commodore, Olivetti, and ACORN (incl. the BBC Micro). Much later, Hell-software was developed for other operating systems such as iOS and Android. Here, I use the word "PC" to denote any personal computing device/platform (desktop, laptop, notebook, tablet, smartphone, ...), independent of manufacturer, operating system, etc.

In 1983 I bought a CHE-1, a cloned-and-improved Apple-II, developed by Computer Hobbyclub Eindhoven. The first Hell-software that I used, ran on this machine. In those days, sofwtare was loaded from a cassette recorder/player! The Hell software was developed by Klaas Robers, PA0KLS (ref. 1). It was a whopping 2 kB in size, and was written in assembler and/or machine language. This Hell-software can be downloaded here (this is an Apple .dsk file that also includes an RTTY and an SSTV program). The screenshots below, and those of the instructions & explanations included as part of the program are here.

hell software

Fig. 1: screen-captures of the PAØKLS Hellschreiber for Apple II

(source right-hand image: ref. 1)

I was able to run this software under Windows® with an Apple-II emulator (ref. 40):

"Apple -Hell - Schreiber" by Klaas Robers (PA0KLS), running on Apple ][ emulator for Windows

(the actual Feld-Hell and Hell-72 screens start about minute into the clip)

At the university, I was still using my "personal" DEC PDP-11/23 at the time (1983-84)! A couple of years later I got an old PC, and found the Hell-software for MS-DOS by Sigfus Jonsson, LA0BX. It can be downloaded here. Onno Hoekstra (PA2OHH) developed a DOS Hellschreiber in 2000 (ref. 6).

A DOS-Command function was available as part of Windows® Win95/98. It actually was a limited DOS-emulator, and it did not provide the required hardware-level access to the LPTx or COM port, even though it may execute the HS.exe program. I installed an excellent DOS x86-emulator (DOSBox, freeware for Windows, Mac OS, OS/2, Linux, etc.). It is geared towards DOS-games. The version that I used (V0.73, May 2009) unfortunately doesn't support LPT ports (the serial COM-port supposedly works). The LA0BX software requires a simple HamCom-type interface as described further below. Like the Apple II program mentioned above, the RX/TX interface is at TTL-level, rather than using a sound-card (though if you crank up the audio output of your RX, you may get these Hell-printers to work without an additional interface).

hell software

Fig. 2: start-up screen and a "help" page of the LA0BX Hell-software for DOS

In those early days of home computing (1980s), Hellschreiber software was also developed for the Z80 (e.g., by Reinier Ijzer, PA3CTL), for ACORN / BBC Micro (e.g., Marien van Westen, PA0MVW; Paul G4BKI/VP9KF), Olivetti M10 / Tandy 100 (e.g., Hans Kreuzer, DL1AN, ref. 4), and Commodore C64 and C128 (Louis Crijns, PE1DXH/PA3DSA, ref. 5). There are C64 emulators for Windows®, e.g., CCS64.

Note that the above Hell-emulators by PAØKLS, LAØBX, and PE1DXH are still the only (non-commercial) ones that include the Hell-72 "GL" start-stop Hellschreiber mode!

Currently, I have several Hellschreiber and Hell-capable multi-mode software programs for Windows® on my laptops. They can all use the PC's sound card or an external USB audio dongle. These days, these audio interfaces are accurate enough to not need a calibration offset. The most popular Hellschreiber-capable programs are:

  • Hellschreiber - freeware from Nino Porcino (IZ8BLY), co-developed with Murray Greenman (ZL1BPU). This is my favorite Hell software. It also appears to be the only one that can automatically save screenshots. I use this feature for my Hell Web-Cam.
  • This wonderful program dates back to the 1990s Win-95 days. You can make it run fine under Win 7/8/10/11, but you may need some help. I have captured this (and general operating hints) in ref. 42.
  • MultiPSK - multi-mode freeware from Patrick, F6CTE. It includes Hell-80 synchronous mode and Hell-80 "start-stop" transmission, and has slant correction buttons during receive.
  • Fldigi - ("Fast Light Digi-mode") freeware by David, W1HKJ. It also includes Hell-80, but synchronous mode (i.e., not start-stop). It runs under Windows, Linux, Unix, and OS X.
  • Digital Master (frmr. DM-780) - very nice multi-mode software by Simon brown (HB9DRV), et al; it is part of the payware package Ham Radio Deluxe (HRD, about $100) that includes a PC-control program for CAT-capable transceivers and some SDRs, antenna rotor control, etc. This was freeware up to  Version 5.24.38, ca. 2012.
  • MixW good multi-mode pay-ware (about €80) by Nick Fedoseev (UT2UZ), and Denis Nechitailov (UU9JDR).
  • WinHell by Henk Spekkink (PA3BQS). This freeware allows the user to define the detector characteristics (curve of signal level vs. pixel brightness). As shown in the screen capture below, this also allows selection of inverse video (white-on-black) which is helpful under noise/interference conditions. It is the only software that includes a character-matrix editor. There is also a beta-version with sliders for the center frequency and bandwidth of the detector.
  • SkySweep was multi-mode pay-ware, discontinued in 2009. It has many configurable signal processing functions, but a rather awkward GUI. It covers Feld-Hell, PSK-HELL-105 and -245.

hell software

Fig. 3: screenshots of the IZ8BLY (left) and DM780 (right) Hellschreiber emulator

hell software

Fig. 4: screenshots of the MixW (left) and Multipsk (right) Hellschreiber emulator

hell software

Fig. 5: screenshots of the Fldigi (left) and WinHell (right) Hellschreiber emulator

hell software

Fig. 6: screenshot of SkySweep Pro 5.11 in Feld-Hell mode

Hell-80 machines can operate in synchronous mode (like the Feld-Hell), and in asynchronous (start-stop) mode like the Hell-72 "GL" and -73 "AGL" machines. The available Hell-80 capable emulators all do the synchronous mode. Only MultiPSK can transmit in the asynchronous mode (i.e., "start-stop"). Unfortunately, Multipsk (and all the others) does not have a start-stop receive mode. Obviously all can receive start-stop transmissions. However, the start-pulse in the first column of each character will simply be printed, rather than be used for start-stop synchronization. So the received characters will dance around on the simulated paper tape...

More exhaustive lists of various Hellschreiber software can be found here (DXzone), here (Ko Versteeg, NL9222), and here (Oliver Welp, DL9QJ).

I have never played around with the Slow Hell mode (the "Hell" counterpart of QRSS CW/Morse). Peter Burri (HB9JAQ) has made a small Windows program "Slow Matrix Transmitter" to send Slow Hell signals. It can be downloaded here. The received signals are directly readable in an appropriately configured waterfall display or spectrum analyzer software (e.g., SpectrumLab, Spectran), or of one of the many "QRSS Grabber" websites.

I primarily use the dedicated Hell-only software, like I do for SSTV and RTTY (MMSSTV and MMTTY by Makoto Mori, JE3HHT). Sometimes I run MixW or DM780 in parallel (I do prefer audio spectrum and waterfall displays in color).

The IZ8BLY can automatically save a capture of the receiver-screen every 2 minutes. I use this feature for my Hellschreiber WebCam. IZ8BLY does not include an ftp feature to automatically upload the screenshots to a website. DM780 does include such an ftp feature, but only for SSTV... I use FTPGetter Pro for automatic uploading of screenshots and WebCam status. FTPgetter is shareware, not freeware; I still use the demo version. My FTPgetter settings (with the exception of the password, hi) are here.

Click here for a write-up by Paul Youn (GØHWC) on how to calibrate the TX/RX clock settings of some of the above software packages. With few exceptions, you have to perform soundcard calibration! And you have to do it separately for each software package that you use! The same applies to SSTV software, by the way.

The above software packages can also be used to test Hellschreiber machines. However, none cover all mechanical Hell modes, in particular Hell "F" (same signal format as "Presse Hell"), synchronous Hell-80, and NVA (ATF) Hell. Ralf Lampe has solved this shortcoming with his EREL-Hell software. Early 2011, he added the NVA (ATF) Hell mode as part of V1.23. It can be downloaded here. Note that this software is explicitly designed for testing Hell machines. The generated tone-pulses are not suitable for modulating a transmitter!

hell software

Fig. 7: start-up screen and GUI of the GUI of the EREL-Hell test-signal generator

At this time (2016), the choice of Hell applications for Linux, iOS/OS X, and Android is more limited:

  • Linux: multi-mode Fldigi (as described above). Apparently, Multipsk can be run under "WINE" on Linux.
  • iOS / OS X: Fldigi, multi-mode cocoaModem by Kok Chen (W7AY; no update since 2012; has slant-correction buttons), and Feld Hellschreiber by Chris Smolinski (N3JLY).
  • Android: Hellschreiber Feld Hell RX/TX freeware (latest version is V4 from 2015). See ref. 7A-7C for Android rig-interfacing issues.

hell software

Fig. 8: screenshot of my old Android tablet (Samsung Tab 3 lite) while receiving Hell signals

Note: there are reports that some of the Android Hell software only runs on older versions of Android (up to version 5, a.k.a. "Lollipop").

My PCs and laptops all use Windows®. With Win7, 8, Vista, and 10, Microsoft has been messing with the very convenient "stereomix" utility. In its usual infinitesimal wisdom and infinite arrogance, Microsoft has decided that its users should no longer have "stereomix". Even worse: starting with Win10, Microsoft has the right to - and will - remove the soundcard driver that includes "stereomix" from your PC, and replace it with their own driver. Extremely annoying when you want to use your digi-mode software when listening to a Web-SDR!

hell software

Fig. 9: Stereomix GUI

An old laptop that I used for amateur radio stuff originally had XP with a Realtek AC67 audio driver. The laptop works fine with Win10, as did the original driver. I managed to re-install the driver (downloaded 32 & 64 bit version R2.81 from Realtek here, 168 MB) and get "stereomix" back. Obviously, each time Microsoft installs a mandatory Win10 update, the generic Microsoft driver is re-installed (this driver update "feature" can be disabled via Control Panel → System → Advanced system settings → Hardware → Device installation settings → No). So I do no have to use an external soundcard or dongle. Some people wrap the loudspeaker/headphone output back to the microphone input (with a potmeter or other simple voltage divider, to reduce the audio level, so as not to overdrive the microphone input). But that is a big hassle!

Fortunately, nowadays there is an excellent alternative to Stereomix": the Voicemeeter virtual audio mixer. Installation is straightforward (but read & follow the instructions!) and after some minor configuration, you are good to go. See ref. 37 for my personal settings file (very easy to import via the Voicemeeter menu). I now use this exclusively with digi-mode software, when I listen to PC-internal audio, such as audio from Web-SDRs. Voicemeeter is professional software and has many nice features, but we basically only need the virtual audio link. Processor load is minimal. Voicemeeter is donationware, which is effectively freeware - with the request to make a donation of your choice, small or big. Please make a donation when you decide to use this software!

hell software

Fig. 10: Voicemeeter GUI

If you'd like to more closely analyze the spectrum of the audio signals you receive (or generate), consider "Spectrum Lab" Windows-freeware from DL4YHF. You can also use this to look at the spectrum of the audio signal that the digi-mode software sends to your transmitter (Spectrum Lab settings to get you going are here). I have put the spectrum plot of a various Hell-modes on the "Modern software" modes (FM, MT, etc.)" page.


Obviously the PC needs to be interfaced with a transmitter/receiver. Typically the following interfaces are required:

  • Audio from the receiver loudspeaker, headphone, or line output to the microphone ("mic") or line-in of the PC's soundcard.
  • Audio from the soundcard to the microphone input of the transmitter.
  • PTT-control from a serial port of the PC to key the PTT-input of the transmitter. You may be able to use VOX instead:
  • Make sure that the "off"-delay is sufficient to cover text spaces.
  • Make sure that you have no other programs running on your PC that can generate sound. You don't want to transmit "You have mail", or alerts from your calendar program, etc.

With original DOS program by Sigfus Jonsson (LAØBX), I used the HAMCOM-style interface designed by Kåre Lind (LA9ZO). See ref. 8. The circuit board for this interface was kindly provided to me by Helmut (DL1OY, SK) in September of 1992.

hell software interfaces

Fig. 11: LA9ZO Hell PC-interface (HamCom-type) - 1990

(top left: I added a transistor for PTT-interface)

To prolong the life of your PC, avoid humms and ground-loops, it is advisable to galvanically isolate the PC and the transceiver from each other - especially if they are not battery powered and you have a transistorized transceiver. This can be done with small isolation transformers for the audio, and opto-couplers for the PTT interface. It is recommended to install the audio isolation even when just using a receiver.

Note: modern transceivers such as the Elecraft K3 and K4, have "Line In" and "Line Out" interfaces that are transformer-isolated and have adjustable signal levels. This greatly simplifies the PC soundcard interface - you only need an audio cable and a PTT interface (or use VOX)!

I have had some problems with RF getting into my PC, and have added ferrite rings (I typically use material 43 for chokes) and clamshell clamp-on ferrites to all AC- and DC-power cables, all audio signal I/O lines, and all USB cables (keyboard, hub, etc.). My old Logitech® cordless mouse also locked-up at some transmit frequencies, and I have reverted to a USB wired mouse. Note that the cordless mouse operated at 27 MHz. Later models run at 2.4 GHz, some at 100-145 kHz; other brands apparently use 13 MHz...

Modern laptops and notebook PCs no longer come equipped with a conventional serial COM-port. This means that the standard interface for the PTT signal is not directly available. USB-RSR232 adapters are available commercially. A USB-RS232 converter chip is readily available: the FT232BM chip from FTDI Ltd. This chip is built into many serial-USB dongles and sticks. It works with all popular digi-mode software packages. A driver for Windows etc. is readily available. Note that some inexpensive dongles/sticks from the Far East have a clone-chip that often has issues with drivers. Caveat emptor... Also, dongles/sticks do not include galvanic isolation (opto-coupler).

The same considerations and precautions apply to Hell modulation as to fast CW, and PC-based digi-modes (ref. 9, 10A-10C, 30). You don't want to transmit a signal that is several kHz wide!

  • Correct setting of the audio level of the PC's sound card line out put.
  • Correct setting of the transceiver's microphone gain.
  • Correct setting of the transceiver's power.
  • Turning off the transmitter's speech processor.
  • Turning off the transmitter's Automatic Level Control (ALC) to avoid clipping (non-linearity).
  • Avoid ground loops (when using a PC and digi-modes software).
  • Avoid picking up 50/60 Hz hum (when using a PC and digi-modes software).
  • Tune the transceiver's frequency, such that the audio frequency of interest is in the middle third of the waterfall spectrum display (typical transceivers have a receive and transmit bandpass of 700-2000 Hz, with rapid fall-off near these band edges).
  • See here for illustrations of the received spectrum of over-driven transmitter and broadband 50/60 Hz and 100/120 Hz artifacts.

Click here for a description of how I hooked up my laptop to my transceiver. Ref. 9, 11-21, 31, 32, 33 provide some more interface descriptions, designs, and considerations.

My Mk II interface

Fig. 12: my 2009 "spaghetti" interface

Note: small, inexpensive audio isolation transformers often have a bandwidth that is narrower than the 300-3300 Hz voice-band specified for conventional phone systems ("POTS"). Example: Radio Shack item 273-1830, shown below. This bandpass will be in series with the receiver and transmitter audio bandwidth of your rig. E.g., my FT-817ND transceiver has a specified TX & RX -6 dB audio response of 400 Hz - 2600 Hz. This is one more reason to operate near 1000 Hz. Note that the original standard Hellschreiber audio frequency was set to 900 Hz in the 1930s, for similar reasons.

My Mk II interface

April 2009: I decided to replace the spaghetti interface with a commercial interface box: a Signalink SL-1+ from Tigertronics. It is about the size of a pack of cigarettes and does the job (well, normally it doesn't smoke, hihi). I did have to reduce the microphone gain setting inside my transceiver, to be able to get Signalink's PTT function to work properly with the PC's Sound Mixer without over-driving the microphone input to the transmitter (this issue is covered by the Signalink instructions). I also had some issues with inadvertent PTT activation at certain combinations of audio source selection ("what you hear" vs. microphone (from RX) and wave (e.g., from web-SDR)) and wave & overall volume setting in the PC audio mixer.


Fig. 13: a commercially sold interface box

December 2009: I have decided to revive my spaghetti interface! Reason: the Signalink Model SL-1+ (and similar type interfaces) do not have input & output volume control potmeters [note that this serious shortcoming has been corrected in the later Signalink models]. For being able to quickly change gain settings, having potmeters is much easier than having to activate and use a soundcard interface utility on a PC. I have put the potmeters and circuitry (now including a PTT-LED) in a small project box, and cleaned up the wiring mess.

My Mk II interface
My Mk II interface

Fig. 14: my 2010 "happy-box"

February 2010: I have expanded the interface with a second output of the PTT-optocoupler. This allows me to direct-key a CW transmitter. The latest circuit diagram is here. A DPDT toggle switch is added to avoid simultaneous activation of the PTT and CW Key lines. Note: currently, only the old DOS and Apple based Hellschreiber software supports direct keying!

December 2012: earlier this year, I acquired a very compact QRP HF/VHF/UHF transceiver, a Yeasu FT-817ND. Obviously I needed a PC interface for this rig as well. I decided to rebuild my existing interface. My new laptop PC doesn't have a serial COM port, so I am using an inexpensive USB-to-serial adapter dongle (Digitus DA-70156) that connects to a 9-pin D-sub connector on my little interface box. Works great, though not via all the USB-hubs that I tried... For the latest version of my straightforward interface circuitry, see ref. 41.

My Mk III interface

Fig. 15: my "third generation" rig-interface "happy box"

My Mk III interface

Fig. 16: the "ugly style" perf board (left) and a top view of the board

My Mk III interface

Fig. 17: my basic interface circuitry

(full schematic: see ref. 41, incl. additional PTT output)

My Mk III interface

Fig. 18: for PC's and laptops without RS-232 COM port or soundcard interface

Note: there are (still) many USB-serial adapter/programming/CAT cables on the market that use the Prolific PL2303 chip. Cheap adapter/programming cables may have a Chinese "pirated" chip (PL2303-H, -HX, -X, -XA, ...). Starting with Windows 7, the official Prolific/Windows drivers no longer work with those counterfeit chips. Also see this Microsoft Community thread and this blog for discussions and solutions. Third-party work-around software appears to be available on the internet (e.g., here), to install driver version (the last compatible version).

The FT-817 transceiver has a 6-pin mini-DIN "data" jack for the audio connections and PTT input. I put such a jack on the back of my interface, so I can use a standard 1:1 male-male cable. I took a second male-male cable and put an 8-pin connector on one end, to plug into my Alinco DX-70TH transceiver. The source of the CW Keying output is selectable with a toggle switch: either the RTS signal of the serial interface (via an opto-isolator), or an open/ground signal that is derived from the audio pulses received from the PC soundcard.

My Mk III interface

Fig. 19: a patch cord for my Alinco DX-70TH (left) and the one for my Yeasu FT-817ND

My station

Fig. 20: my fixed-base station with all its interfaces (2015)

I also have a tiny Baofeng UV-3R handheld VHF/UHF transceiver "toy" (but it works). So, obviously I needed a PC-interface for this one as well. I kept it very simple: just 2 trim-pots for the RX and TX audio. No need for a PTT line, as the UV-3R has a VOX function. No galvanic isolation with audio-transformers either. For simplicity, and because the handheld transceiver runs off its own battery. When I also acquired a UV-5R handheld, I had to change the connector on the transceiver side of the interface:

baofeng interface

Fig. 21: schematic of my simple audio interface for the UV-3R and UV-5R

baofeng interface

Fig. 22: inside are just two 50 kΏ trim-pots, glued to the inside of the 3.5x5 cm project box

I cannibalized the earphone/microphone of a portable telephone to get the 4-conductor 3.5 mm plug that is required for the UV-3R. I basically cut out the microphone, re-used the cables/wires and replaced the earbuds with 3.5 mm stereo plugs. For the UV-5R, I bought a similar earphone/microphone with the right connector on eBay for less than $2 (2015).

baofeng interface

Fig. 23: earbud headphone with microphone and 4-pole plug

Hellschreiber and other digi-mode software is available for tablets and smartphones. Interfacing them with a transceiver is basically the same as with a PC. Interfaces are available commercially (e.g., ref. 34, incl. schematic), but they are just as easy to make yourself, as for a PC. The required 4-pole 3.5mm plug - with wires - can be taken from an inexpensive extension cable, or from earphones, as described above (Fig. 23). Simple VOX-circuitry is required to generate the PTT signal for the transmitter. A possible interface circuit with VOX/PTT is shown below.

tablet smartphone interface

Fig. 24: simple transceiver interface for tablets and smartphones

(source: adapted from ref. 35)

Some notes regarding the schematic:

  • The circuit does not require a supply voltage.
  • The 4.7 kΩ resistor in series with the 1:1 audio isolation transformer is needed, in order to enable the "external microphone detected" function in the tablet/smartphone to work.
  • Many transceivers have a microphone input with a bias voltage (e.g., for the electret microphone of headsets). The 10 μF capacitor at the bottom of the circuit schematic blocks this voltage. The capacitor may be omitted, if the transceiver does not apply such a bias voltage.
  • There is a separate audio transformer, to make the VOX function faster and more sensitive ( = requiring lower volume setting on the tablet or smartphone). That audio transformer should have an impedance transformation ratio of about 1:5 or more. I.e., a voltage transformation ratio of at least 1:2. This is easy to obtain with a 1:1 transformer that has a center tap.
  • The transistor should have a large current gain hFE of 300-400. The diodes should have a small forward voltage drop.
  • The pin-out of the 4-pole plug is for tablets and smartphones that are compliant with the CTIA standard.

More and more, tablets and smartphones no longer have a 3.5 mm audio jack. Instead, they have a USB-C jack (24 pins). USB-C does NOT include an analog audio interface! However, a large variety of simple and relatively inexpensive USB-C/3.5mm adapters is on the market. They have built-in D-to-A & A-to-D conversion hardware. They can be used to hook up an analog audio device such as an analog headset with microphone, analog earbuds with microphone, a transceiver interface,...

tablet smartphone interface

Fig. 25: An adapter cable from USB-C to 3.5 mm analog audio - with ADC/DAC circuitry built into the male USB connector

(note: the pin-out of the USB-C connector is rotationally symmetrical: the connector also works upside-down)

Now suppose you want to use a Hellschreiber or other digi-mode application on your tablet or smartphone to "print" the signals from an on-line receiver, such as a Web-SDR. With Android, you will not be able to do so. So your only option is to loopback the signals from the headphone/earphone output back to the microphone input. This wrap-around is quite easy to implement with just a few resistors and a small capacitor. This is a 4-wire interface: 2x audio output (left, right), 1x microphone input (mono), and ground/common. Note that this interface is also used by the tablet or smartphone to detect presence of headphones and microphone, and to input commands for volume up/down, play/pause, etc. So the component values must be chosen with this in mind. Also, the resistors values must be chosen so as to reduce the microphone input level, such that it is not overloaded by the looped-back audio:

tablet smartphone interface

Fig. 26: CTIA/TRRS headphone/microphone interface with audio loopback

(source: adapted from ref. 38)

tablet smartphone interface

Fig. 27: The loopback dongle plus USB-C to CTIA/TRRS adapter cable for my Android tablet and Android smartphone

This dongle works fine with my old Android tablet (a Samsung Galaxy Tab3 Lite) and with my Android smartphone (an OPPO Find X2 Neo, via a USB-C to 4-pin 3.5mm CTIA adapter cable, see Fig. 25).

BE CAREFUL: there are two Tip-Ring-Ring-Sleeve (TRRS) pin-outs - CTIA (Cellular Telecommunications Industry Association) and OMPT (Open Mobile Terminal Platform). They have opposite "microphone" and "ground" wiring! "CTIA" has microphone/AUX connected to the sleeve (i.e., farthest from the tip) and Ground to the ring next to the sleeve. OMPT has Ground connected to the sleeve, and microphone/AUX to the ring next to the sleeve.

The next diagram shows another digi-mode interface for smartphones and Android tablets, designed by Christian Petersen (DD7LP, ref. 39). Not all smartphones and tablets are created equal. E.g., microphone input impedance may vary. For a Samsung Galaxy S3, resistors R1 ands R7 (both 33 kohm nominal) may have to be reduced to 8.2 or 10 kohm, and both C5 and C6 reduced to 4.7 uF.

tablet smartphone interface

Fig. 28: DD7LP Smartphone / Android tablet  interface to a Yeasu 817 transceiver

(source: adapted from ref. 39, 2013)

Wolfgang Philipps (W8DA) uses a similar schematic (note: there are differences!) for the small interface box that he sells (ref. 34):

tablet smartphone interface

Fig. 29: Wolphi (W8DA) Smartphone / Android tablet interface to a Yeasu 817 etc. transceiver

(source: ref. 34)

In the above two schematics, some folks have successfully used a 2N2222A transistor for T1 and a 2N7000 FET for T2.


Does the waterfall display of your digi-mode software have a lot of 50/60 Hz harmonics and other noise in it - that doesn't come from your receiver or antenna? Chances are, if you see such noise and harmonics garbage in your waterfall display, it may also be on the audio output of your PC! This means that you may actually be transmitting it with your digi-mode software. You don't want that!

The screenshot below shows what my 100-2500 Hz waterfall display used to look like - not anymore! For a long time, I blamed it on QRM generated in/by the apartment building that I live in, and big-city QRM levels in general.

equipment bonding

Fig. 30: lots of noise and 50 Hz power harmonics in my receiver's audio spectrum

All the way up to about 8-9 MHz, my receiver's S-meter showed a noise level of at least S8 - try working DX with that!

equipment bonding

Fig. 31: S-meter without bonding wire (noise = S8)

Then I installed a multi-strand copper bonding wire (#14 AWG, 1.6 mm Ø) between the chassis of my laptop-PC and the ground/return of the external 12 VDC power supply of my transceiver. My laptop is an older model Dell, so it has several D-sub connectors: a DB9 serial-port, and a DB15 display output. Their shells are connected to the laptop's chassis. The metal shell of some other connectors (USB, FireWire) is also connected to the chassis. The DB-connectors have two nuts, for the lock-screws of the mating plug. I made a bonding wire with a small ring lug on one end, and a large ring lug on the other. The small lug is screwed onto one of the D-sub connectors. The large lug goes on the "-" binding post of my transceiver's external 12 VDC power supply.

equipment bonding

Fig. 32: the bonding wire is connected to the chassis ground of my laptop

This made quite a difference!

equipment bonding

Fig. 33: my receiver's audio spectrum - after installing a bonding wire between PC-chassis and 12 VDC power ground

To determine where the remaining noise traces came from, I disconnected the laptop's external power adapter (from the wall outlet!!!) and ran the laptop on its internal battery. Ref. 22A. The audio spectrum is now basically clean above about 300 Hz:

equipment bonding

Fig. 34: installing the bonding wire and running on battery completely cleaned up the audio spectrum

On 80m, the S-meter of my receiver also shows the noise reduction: as much as 5 S-points!

equipment bonding

Fig. 35: S-meter with laptop on battery (power adapter disconnected) and with bonding wire (noise = S2+)

equipment bonding

Fig. 36: comparison of audio spectrum with laptop on battery vs. laptop on external power adapter

(bonding wire installed in both cases)

The particular laptop power-adapter that I used with this laptop, is a cheap, simplistic switched power supply (not the original OEM adapter). It has totally inadequate filtering on both the input side (ref. 22B, 22C) and the output side (ref. 22C). Note that not all laptop power adapters (or desktop internal power supplies) are this bad!

equipment bonding

Fig. 37: the circuit card of a cheap laptop power adapter

Note that interference from the power adapter is not just conducted to the transceiver. It is also radiated. If I unplug the 19 VDC from the laptop and touch the barrel connector (and not touch anything else), most of the power supply noise returns!

The next figure shows how to properly suppress the electrical noise that is created by an insufficiently filtered switched power supply. As an absolute minimum, a line filter should be used.

noise supression

Fig. 38: suppression of laptop power supply noise

(source: adapted from ref. 22D)

The material of the ferrite toroidial core must be appropriate for the switching-frequency of the power supply. This is typically 10's of kHz, in which case ferrite material 26 (manufatured Fair-Rite, resold by Micrometals and Amidon) is suitable. A differential-mode filter of 2x 16 turns on an FT-106-26 ring (yellow-white) should be sufficient (but note the opposite winding directions - see the figure above!).

I have a second laptop, on the same desk as the other one. It is not connected to my transceiver, other than via the filtered 220 VAC power strip. I have now also added a bonding wire between this second laptop and the power-supply of my transceiver. This has reduced the receiver noise a tiny bit more. Every little bit helps! As this laptop is in a docking station, the bonding wire is screwed onto one of the D-sub connectors of the dock.

equipment bonding

Fig. 39: S-meter when second laptop is now also bonded to the power-supply of the transceiver

NOTE: the added bonding wires may defeat the galvanic isolation in the PC-transceiver interface circuitry! In my system, this has not caused any problems.


Note: the modern Hellschreiber software discussed above, do not support direct keying of a CW transmitter! Only the old DOS-based and Apple II Hellschreiber software supported direct keying! The only option is to use a keying circuit between the audio output of the PC's soundcard, and the keying input of the CW rig. The keying-circuit recovers the original binary (DC) pulses, by rectifying and filtering the tone pulses. The recovered pulses are passed to a keying tube (or transistor or solid/state relay), the output of which is connected to the keying input of a CW transmitter.

  • The Hell Co. designed tube-based keying devices to do exactly that (1940). See figure 14 in ref. 23.
  • For use with a solid-state CW transmitter, a keying circuit can be made by expanding a CW/Hell demodulator/detector (ref. 24, 25A, 25B) with a keying transistor (or an OptoMOS opto-coupler solid-state relay, ref. 36).
  • Likewise, VOX circuitry can be used, e.g. one of the circuits by Andy Talbot (G4JNT), ref. 26.
  • A keying circuit can be as simple as a diode rectifier, an RC-filter, and a switching transistor. Two diodes and two capacitors can be configured as a full-wave voltage-doubler, that consists of two half-wave rectifiers operating on alternating polarities (ref. 25).

So you want to key a CW transmitter (QRP or other) with the tone pulses generated by your Hell (or CW) software? Unless you are using DOS or Apple II software (see above on this page), you will have to build some hardware to recover the DC/binary pulses from the tone pulses: modern operating systems (Windows, Linux,..) cannot and do not support proper real-time control of outputs!

In 1980, Klaas Robers (PA0KLS), was one of the first to publish a schematic for this purpose (ref. 25A, 25B). It uses a small audio transformer with a center-tap, to increase the signal level. RadioShack used to sell them (catalog # 273-1380, $2.99 2016 pricing). The up-transformed signal is passed through a full-wave rectifier, and amplified by a general-purpose NPN transistor. When the input signal level is sufficient, pulses appear at the output of the transistor. These are used to trigger a retriggerable monostable multivibrator. The time-constant of the multivib has been chosen such that the output is constantly active for input tones above 500 Hz. The schematic below shows the circuit from Klaas, expanded with an opto-coupler. If you are not worried about galvanic isolation and voltage difference between the transmitter and the +5 Vdc supply, then you can leave out the opto-coupler and key directly with the inverted-Q output of the 74123. The 74123 can sink 20mA, the 74HC123 25 mA. In this case, connect the LED to the non-inverting Q output via a 270 ohm resistor (as in the original circuit).

cw keying

Fig. 40: CW/Hell interface for keying a CW-transmitter by Klaas Robers (PA0KLS), expanded with an opto-coupler

(source: adapted from ref. 25)

Max Perner (DM2AUO), has modernized this concept: instead of the transformer/rectifier and transistor amplifier, he uses a high-gain OpAmp to convert a sine-wave at the input into a square wave. Here too, the output of the amplifier goes to a retriggerable monostable multivib. Ref. 26. The circuit is dimensioned to recover pulses with a tone-frequency of 500-1500 Hz. As always, it is best to keep the tone frequency in the middle of the passband of the transceiver (e.g., 1000 Hz). The required +5 Vdc supply voltage is taken from a USB port.

cw keying

Fig. 41: CW/Hell interface for keying a CW-transmitter by Max Perner (DM2AUO)

(source: ref. 28)

I have built this circuit, but had to increase the time-constant capacitor (between pin 14 of the IC and ground) from the original 45 nF value to 150 nF, and used a smaller resistor value for the power-on LED. For experimentation purposes, I also included the adjustable pulse-extender circuitry described further below (the 14-pin IC at the bottom of the photo in Fig. 38).

cw keying

Fig. 42: the above interface circuit, built by me in May of 2013

The oscilloscope screen captures below show the output of the detector (without the pulse extender) for the leading and trailing-edge of a 1000 Hz raised-cosine pulse.

cw keying

Fig. 43: output of the detector for the leading and trailing edge of a 1000 Hz raised-cosine tone pulse

cw keying

Fig. 44: detector output for a raised-cosine pulse

Some modern transceivers do not like (very) high-speed CW-keying or Hell-keying. They limit the rise- and fall-time of keying pulses, often to at least 2 msec. The result is that length of the transmitted pulses is reduced. Also, most of the modern Hell-capable digi-mode software apply raised-cosine pulse shaping. As a result, the pulse-recovery circuitry always loses some of the pulse duration. Feld-Hell has a shortest black or white pixel pulse duration of 8.16 msec (unless you are one of those clowns that uses "nice" Windows fonts, instead of the prescribed Hell-font: shame on you!!!). With a standard 1000 Hz tone, 8.16 msec is only about 8 cycles of the sine-wave. So, losing even a single  cycle is significant.

Years ago, Andrew Lovell (SM6MOJ), and Marien van Westen (PA0MVW) came up with the following circuitry to add some adjustable "hang-time" to the trailing-edge of the detected pulses. This pulse-stretcher circuit uses one 4011 quad dual-gate NAND. Note that the 7400 quad NAND has a different pin-out! The 10 kΩ trimpot allows the input pulse to be extended by as much as 25 msec. We actually don't need that much, so a smaller trimpot may be used.

cw keying

Fig. 45: pulse-length extender by Andrew Lovell (SM6MOJ) and Marien van Westen (PA0MVW)

In the above detector/keyer circuits (Fig. 40, 41), the pulse-extender can be placed between the inverted-Q output of the multivib and the opto-coupler:

cw keying

Fig. 46: pulse-detector combined with the pulse-extender

cw keying

Fig. 47: pulse-extender signals - left: pins 4/13 (top), pin 10/12 (bottom); right: pins 8/9 (top), pin 10/12 (bottom)

(the pin numbering refers to Fig. 45)

cw keying

Fig. 48: pulse-extender signals - input (top) and output (bottom)

The above circuits all require a supply voltage. The following circuit shows how you can do without! The tone pulses are passed through a simple 1:1 audio isolation transformer and then to a regular diode/capacitor voltage-doubler. The rectified signal controls a switching-transistor. The audio transformers can often be salvaged from discarded old modem-cards. Instead of the NPN transistor, you could also use a self-latching FET, such as the BSN10A. 1N4148 diodes may be used instead of the 1N914. Use rectifier diodes with a small forward voltage, i.e., not silicon diodes such as 1N4148, 1N914, BAV21. Instead, use AA118 (best), 1N34A, 1N5711, BAR28, or similar.

cw keying

Fig. 49: CW/Hell interface for keying a CW-transmitter by Mike Blake (K9JRI)

(source: ref. 27)

In a variation on the above schematic, 1 nF capacitors are added (from both transformer output pins to ground), and the upper 10 kOhm resistor is reduced to 1 kOhm.

The lower half of Fig. 24 above is another PTT circuit. The following simple circuit has been used by some folks:

cw keying

Fig. 50: another CW/Hell interface for keying a CW-transmitter

Note: the above interfaces are only intended for use with solid-state transmitters! Tube transmitters typically have high positive or negative voltage at the keying input - as much as several 100 volt! If you want to key such a transmitter with one of the above pulse-detectors, you must add a (simple) high-voltage keyer interface. An excellent one is described in ref. 18 (incl. schematic; also available as a kit).

Note: there are two basic forms of keying in tube/valve transmitters: grid-block keying (changing either the grid bias or the cathode bias voltage), and anode keying. Keying the anode voltage requires the keying circuitry (or the hand key, as the case may be) to handle both high voltages and high currents! This is rather hard to implement with keying relays (conventional relays, reed relays, mercury-wetted relays).

Make sure that your CW transceiver is suitable for high-speed Morse code! Feld-Hell has an equivalent telegraphy speed of 30 WMP (2.5 chars/sec), and a shortest pulse duration of 8.16 msec. If your transceiver has a "full break-in" option, do not activate it! It is often implemented with a small relay, that will be "chattering" for every transmitted pixel.

Obviously, the problem of keying a CW-transmitter with Hell pulses is nothing new. Rudolf Hell and his co-workers (ref. 23, 29) recommended direct keying with a Feld-Hellschreiber machine only for low power tube transmitters. For medium power (up to 100 W) they describe a keying device ("Tastgerät") between the drum contacts and the transmitter's key input. The proposed keying circuit takes the continuous 900 Hz tone of the Feld-Hellschreiber (available at the 12-pin round connector of the Feldfernschreiber), transforms its 2 Vpp up to 400 Vpp, and uses a full-wave rectifier tube to rectify this continuous tone. Ripple smoothing is done with a simple RC-filter. The resulting high DC-voltage is passed through a current-limiting resistor, and used as the grid-blocking voltage of a transmitter's modulator or final stage. The Hellschreiber's character-drum contacts (also accessible at the 12-pin connector) are connected across that key input. They short the blocking voltage to ground when pixels are transmitted. Anode and filament heater power for the rectifier tube is also available at the 12-pin connector.

I have not (yet) been able to determine if Hell or Siemens-Halske ever manufactured such keying devices.

cw keying

Fig. 51: Rudolf Hell's keying circuit for medium power CW transmitter

(adapted from figure 13 in ref. 23, figure 7 in ref. 29)

For high power transmitters, Hell et al suggest a keying circuit that uses the keyed 900 Hz tone from the Hellschreiber, taken directly from the La-Lb/E phone line connector of the Feld-Hellschreiber, or even from that connector to a remote transmitter via phone lines. The keying circuit is similar to the one that is used in the Hellschreiber itself for "keying" the electro-magnet solenoid of the printer: a transformer coupled pre-amplifier, followed by a transformer-coupled rectifier with RC-filter, and a keying tube.

cw keying

Fig. 52: Rudolf Hell's keying circuit for high power CW transmitter

(adapted from figure 14 in ref. 23; schematic shows simplified rectifier)


Note 1: due to copyright reasons, this file is in a password-protected directory. Contact me if you need access for research or personal study purposes.

External links last checked: December 2019, unless noted otherwise

red-blue line